Although most deviations from trauma resuscitation protocols are variations that result from the flexibility needed for managing patients with differet injuries, other deviations are errors that can contribute to significant adverse patient outcomes Our long-term goal is to develop computerized decision support for trauma resuscitation and other fast-paced, high-risk critical care settings that monitors workflow for deviations that are known to be associated with adverse outcomes and provides alerts to these deviations, allowing remedial actions to be taken to prevent adverse outcomes. The overall objectives for this proposal, which are the next steps in the attainment of this long-term goal, are to: (a) develop a scalable approach for recognizing activities during trauma resuscitation; and (b) identify deviations associated with adverse outcomes within the workflow of trauma resuscitation using process mining. The central hypothesis is that trauma resuscitation activities can be monitored and analyzed in real time for workflow deviations that increase the likelihood of adverse patient outcomes. The rationale for the proposed research is that real-time identification of risk conditions for adverse outcomes will allow medical teams to take measures for reducing or preventing the impact of medical errors. The central hypothesis will be tested by pursuing two specific aims: 1) develop a scalable and automatic approach for creating an event log of activities occurring during trauma resuscitation; and 2) identify and characterize the team's ability to manage major errors during trauma resuscitation. Under the first aim, the approach will involve (i) the use of radiofrequency identification (RFID) technology and other modalities to create resuscitation event logs of human movement and object use and (ii) comparisons of sensor logs with logs obtained using manual video review (ground truth). For the second aim, the approach will involve the development and refinement of knowledge-based resuscitation workflow models using consensus sequences of activities from manually captured event logs. This project is significant because these methods are an essential early step toward the development of computerized decision support systems that can improve outcomes by monitoring and supporting the work of critical care teams. The proposed research is innovative because it represents a substantive departure from the status quo, focusing on developing methods for obtaining data from sensors to automatically track multiple, concurrent activities and for detecting deviations associated with adverse outcomes within a variable workflow. These methods are expected to form a basis for computerized systems for real-time decision support of medical teams that improve patient outcome during trauma resuscitation and other critical care processes.